Google search is losing the fight with SEO spam, study says

It's not just you—Google Search is getting worse. A new study from Leipzig University, Bauhaus-University Weimar, and the Center for Scalable Data Analytics and Artificial Intelligence looked at Google search quality for a year and found the company is losing the war against SEO spam.

The study, first spotted by 404media, "monitored Google, Bing and DuckDuckGo for a year on 7,392 product review queries," using queries like "best headphones" to study search results. The focus was on product review queries because the researchers felt those searches were "particularly vulnerable to affiliate marketing due to its inherent conflict of interest between users, search providers, and content providers."

Overall, the study found that "the majority of high-ranking product reviews in the result pages of commercial search engines (SERPs) use affiliate marketing, and significant amounts are outright SEO product review spam." Search engines occasionally update their ranking algorithms to try to combat spam, but the study found that "search engines seem to lose the cat-and-mouse game that is SEO spam" and that there are "strong correlations between search engine rankings and affiliate marketing, as well as a trend toward simplified, repetitive, and potentially AI-generated content."

The study found "an inverse relationship between a page’s optimization level and its perceived expertise, indicating that SEO may hurt at least subjective page quality." Google and its treatment of pages is the primary force behind what does and doesn't count as SEO, and to say Google's guidelines reduce subjective page quality is a strike against Google's entire ranking algorithm.

The bad news is that it doesn't seem like this will get better any time soon. The study points out generative AI sites one or two times, but that was only in the past year. The elephant in the room is that generative AI is starting to be able to completely automate the processes of SEO spam. Some AI content farms can scan a human-written site, use it for "training data," rewrite it slightly, and then stave off the actual humans with more aggressive SEO tactics. There are already people bragging about doing AI-powered "SEO heists" on X (formerly Twitter). The New York Times is taking OpenAI to court for copyright infringement, and a class-action suit for book publishers calls ChatGPT and LLaMA (Large Language Model Meta AI) "industrial-strength plagiarists." Artists are in the same boat from tools like Midjourney and Stable Diffusion. Most websites do not have the legal capacity to take on an infinite wave of automated spam sites enabled by these tools. Google's policy is to not penalize AI-generated content in its search results.